A Tool for Model Generation and Knowledge Acquisition

Tools to automatically induce domain descriptions from examples are valuable aids for the knowledge acquisition stage of expert system construction. This paper presents a description of an algorithm that induces two domain descriptions: a conceptual model, which gives a broad understanding of variable interactions and their effect on the system, and a predictive model, which determines the system value associated with variable values input to the model. This induction algorithm is based on Entropy Data Analysis (EDA), which builds linear equations to approximate the system described by the training data.